paper-with-me

홈 › Papers

Preference-Guided Reflective Sampling for Aligning Language Models

2024-08-22 · Hai Ye, Hwee Tou Ng

Iterative data generation and model re-training can effectively align large language models(LLMs) to human preferences. The process of data sampling is crucial, as it significantly influences the success of policy improvement. Repeated random sampling is a widely used method that independently queries the model multiple times to generate outputs. In this work, we propose a more effective sampling method, named Preference-Guided Reflective Sampling (PRS). Unlike random sampling, PRS employs a tree-based generation framework to enable more efficient sampling. It leverages adaptive self-refinement techniques to better explore the sampling space. By specifying user preferences in natural language, PRS can further optimize response generation according to these preferences. As a result, PRS can align models to diverse user preferences. Our experiments demonstrate that PRS generates higher-quality responses with significantly higher rewards. On AlpacaEval and Arena-Hard, PRS substantially outperforms repeated random sampling in best-of-$N$ sampling. Moreover, PRS shows strong performance when applied in iterative offline RL training.

📄 PDF Abstract BibTeX arXiv:2408.12163

Code (1)

nusnlp/prs 공식 구현 pytorch

Tasks

Document SummarizationInstruction FollowingOffline RLResponse GenerationText Generation

Methods 이 논문이 사용한 방법론

ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

Reflective Preference Optimization (RPO): Enhancing On-Policy Alignment via Hint-Guided Reflection

2025-12-15 · Zihui Zhao, Zechang Li arxiv

Direct Preference Optimization (DPO) has emerged as a lightweight and effective alternative to Reinforcement Learning from Human Feedback (RLHF) and Reinforcement Learning with AI Feedback (RLAIF) for aligning large lang…

Reinforcement Learning

Clear Preferences Leave Traces: Reference Model-Guided Sampling for Preference Learning

2025-01-25 · Nirav Diwan, Tolga Ergen, Dongsub Shim, Honglak Lee

Direct Preference Optimization (DPO) has emerged as a de-facto approach for aligning language models with human preferences. Recent work has shown DPO's effectiveness relies on training data quality. In particular, clear…

Math

Reinforcement Learning from Reflective Feedback (RLRF): Aligning and Improving LLMs via Fine-Grained Self-Reflection

2024-03-21 · Kyungjae Lee, Dasol Hwang, Sunghyun Park, Youngsoo Jang 외

Despite the promise of RLHF in aligning LLMs with human preferences, it often leads to superficial alignment, prioritizing stylistic changes over improving downstream performance of LLMs. Underspecified preferences could…

Mathematical Reasoning

LLMdoctor: Token-Level Flow-Guided Preference Optimization for Efficient Test-Time Alignment of Large Language Models

2026-01-15 · Tiesunlong Shen, Rui Mao, Jin Wang, Heming Sun 외 arxiv

Aligning Large Language Models (LLMs) with human preferences is critical, yet traditional fine-tuning methods are computationally expensive and inflexible. While test-time alignment offers a promising alternative, existi…

WildFeedback: Aligning LLMs With In-situ User Interactions And Feedback

2024-08-28 · Taiwei Shi, Zhuoer Wang, Longqi Yang, Ying-Chun Lin 외

As large language models (LLMs) continue to advance, aligning these models with human preferences has emerged as a critical challenge. Traditional alignment methods, relying on human or LLM annotated datasets, are limite…